A Review of Critical Features and General Issues of Freely Available mHealth Apps For Dietary Assessment
←
→
Page content transcription
If your browser does not render page correctly, please read the page content below
Transportation Research Record 2020, Vol. XX(X) 1–?? A Review of Critical Features and General ©National Academy of Sciences: Transportation Research Board 2020 Issues of Freely Available mHealth Apps Article reuse guidelines: sagepub.com/journals-permissions For Dietary Assessment DOI: 10.1177/ToBeAssigned journals.sagepub.com/home/trr SAGE Ghalib Ahmed Tahir1 , Chu Kiong Loo1 , Nadine Kong 2 and Foong Ming Moy 2 Abstract Obesity is known to lower the quality of life substantially. It is often associated with increased chances of non- arXiv:2008.09883v4 [cs.CY] 11 Jul 2021 communicable diseases such as diabetes, cardiovascular problems, various cancers, etc. Evidence suggests that diet- related mobile applications play a vital role in assisting individuals in making healthier choices and keeping track of food intake. However, due to an abundance of similar applications, it becomes pertinent to evaluate each of them in terms of functionality, usability, and possible design issues to truly determine state-of-the-art solutions for the future. Since these applications involve implementing multiple user requirements and recommendations from different dietitians, the evaluation becomes quite complex. Therefore, this study aims to review existing dietary applications at length to highlight key features and problems that enhance or undermine an application’s usability. For this purpose, we have examined the published literature from various scientific databases of the PUBMED, CINAHL (January 2010-December 2019) and Science Direct (2010-2019). We followed PRISMA guidelines, and out of our findings, fifty-six primary studies met our inclusion criteria after identification, screening, eligibility and full-text evaluation. We analyzed 35 apps from the selected studies and extracted the data of each of the identified apps. Most of the apps are engaging, according to user feedback (68%). 62% of the apps provide timely alerts to the user, and 53% of survey apps include goal-settings features. We indicated existing apps are lagging in several aspects. Only 37% of the survey application have included validated databases, 22% of the surveyed applications have addressed data privacy issues, and three applications out of 35 provide offline access to the user. Following our detailed analysis on the comprehensiveness of freely available mHealth applications, we specified potential future research challenges and stated recommendations to help grow clinically accurate diet-related applications. Despite considerable advancements in medicine today, the for Android and iOS devices (2) increased from 640 million number of people getting affected by chronic diseases is to 2,562 million in 2016 (3). These days, smartphone appli- significantly greater due to unhealthy lifestyles. Obesity is cations are being extended to support electronic healthcare one of the most common contributing factors to chronic practices (4), (5), and evidence across several fields show diseases, affecting almost every part of the world from middle promising results which support the feasibility, acceptability to lower-income countries. According to a survey in 2016, 1.9 and efficacy of digital health interventions in different med- billion adults aged 18 years and older were overweight (1). ical conditions. These conditions include but not limited to The prevalence of the aforementioned diseases poses serious managing adolescent health and wellness (92), interventions concerns. However, determining the right remedial measures in sickle cell disease (94), pediatric cancer (93) (97), chronic is dependent on different factors ranging from a person’s health conditions (95) and improving adherence to preventive genetics to lifestyle, which need to be adjusted according behaviour (96). Similarly, e-Health and related diet-related to the cause and severity of the condition. Treatment may applications are being increasingly used for professional include medication, lifestyle changes (87) such as choosing healthier food alternatives, exercise, and requiring patients to follow a customized diet plan (88). 1 Department of Artificial Intelligence, Faculty of Computer Science and Information Technology, University of Malaya, Kuala Lumpur, Malaysia 2 Julius Centre University of Malaya, Department of Social and Preventive On the other hand, with rapid technological advancements and increased usage of handheld devices such as smart- Medicine, Faculty of Medicine, University of Malaya, Kuala Lumpur, Malaysia phones, tablets, and smartwatches, people’s reliance on these devices has undoubtedly grown beyond their utility as a Corresponding author: means to communicate. The number of mobile users in 2012 Prepared using TRR.cls [Version: 2020/08/31 v1.00]
2 Transportation Research Record XX(X) and personal purposes. People are using these applications supports the effectiveness of digital interventions can learn a to make healthier lifestyle choices. Generally, these apps valuable lesson from the findings of this study. provide instant nutritional values of food items with barcode scanners, which is extremely helpful for people suffering Methods from non-communicable diseases, and others who intend to choose healthier products (1). These applications not only We have developed the review protocol by defining assist users with the selection of more nutritious alternatives our research questions and considering multiple inclu- but also allow them to self-monitor their physical activity and sion/exclusion criteria. Then we formally defined our search diet intake by using behavioural strategies of goal settings strategy by identifying the search terms and carried out the (6). Moreover, these applications are designed to cater to search using the electronic database of PUBMED, CINAHIL, various age groups, including children battling obesity from and Science Direct. Following this, we selected relevant stud- a very young age. ies based on our study selection criteria. Then we extracted In this regard, recent developments in artificial the data and presented our results. intelligence-based functionalities and hardware capacity enhancement of handheld devices have led to the Research Questions development of automatic food recognition and calories The primary aim of this review is to answer the research estimation methods, making them an essential subset of questions shown in Table 1. e-health applications. Regardless of numerous diet-related applications freely available today, scientifically proven guidelines (both in Inclusion/Exclusion Criteria terms of usability and functionality) have not yet emerged The studies that met all of the following criteria are selected from the users’ and the dietitian’s perspective. Also, the for this review. author’s first-hand app development experience (7) suggested IC1. Papers related to dietary applications for smartphones a dire need to have in-depth knowledge about the state-of-the- (iPhones, Android phones, and Blackberries) and modern art diet-related applications. To develop this understanding, commercially available portable devices such as iPads and the first step involves identifying key components relevant Personal Digital assistants (PDAs). to existing diet-related applications, which are categorized IC2. Content is written in the English language only. in terms of general issues faced by dieticians and users, IC3. The study must be a full peer-reviewed paper (not an including user experience of both parties and functionalities abstract). required by each of them, respectively. IC4. Dates of Publication: PubMed and CINAHL: 1/1/10- While the development of diet-related applications 31/12/19, Science Direct: 2010-2019 requires a significant amount of time and effort, general This review excludes the following studies that are issues like their credibility remain a question. The term conformed to at least one of the following criteria. "Credibility" here refers to the authenticity or scientific EC1. Studies without a clear description of dietary validation of an application to achieve goal during application mentioned. trials. Another challenge present applications face is the EC2. Dietary applications that are not freely available. maintenance of an updated food composition database, as new food products are being continuously introduced in the Database Identification market. Mobile app developers also find it challenging to determine target users, their needs, and potential feedback We choose PubMed, CINAHL and Science Direct due to to improve functionality and usability of apps (7) (8). Thus, the following reasoning. PubMed database gives a publicly an application with a good user experience may increase its available search interface for MEDLINE and National preference over others despite offering lesser functionality. Library of Medicine, which makes it one of the most widely Therefore, the development of such applications should accessible biomedical resources globally (89). Similarly, the strongly consider essential factors like usability and ’ease of CINAHL database provides allied health care literature, thus use’ (9) (10), as poor usability can result in users switching making it a good resource for literature related to mHealth over to alternative options (11) (7) (12). applications (90). We selected Science Direct as it provides broad access to a database of scientific and medical research The following paper aims to provide a review of existing (91). diet-related applications and seeks to equip researchers and dietitians with comprehensive knowledge about general issues encountered by their users in terms of usability and Search Strategy functionality. Thereby laying the foundation for developing We have carefully defined the search terms based on initial state-of-the-art generalized solutions that can cater to vastly screening through a consensus among authors to investigate varying user needs. Moreover, other fields in which evidence the diet-related mobile applications. Terms such as cellular Prepared using TRR.cls
Ghalib Ahmed Tahir, Chu Kiong Loo, Nadine Kong and Foong Ming Moy 3 Table 1. Research Questions No Research question Motivation RQ1 What are the general problems that are resolved by To provide information about the general prob- the freely available diet-related applications? lems faced by dietary assessment apps such as frequent app crashes (82), cumbersome pro- cess of entering meal information, demotivating information displays (50), periodic notifications, difficulty in portion size estimation (80), credi- bility (78, 84), etc. RQ2 What are back-end application issues resolved by To provide information about the application’s the freely available dietary applications? stability, usage reports (104), data confidential- ity (85), and offline accessibility-related issues faced by diet-related applications. RQ3 To what extent do the freely available dietary To provide information regarding the critical applications fulfill user interface requirements? user interface components (28, 74–77) catered by diet-related applications. RQ4 What are the dietary components and critical To determine and provide information regarding features implemented by the freely available dietary the dietary components (78, 79, 84) and critical applications? features implemented by existing diet-related applications. RQ5 What are the benefits and challenges stemming To summarize the benefits of dietary-related from the included case studies? apps and the challenges they face based on the included studies. phone, mobile phone, smartphone, mHealth, iPads combined with terms like diet, food and nutrition are qualified as keywords in our work. In PubMed, we limited the search to research articles of Clinical Trial, Meta-Analysis and Randomized Controlled Trial published between 1st January 2010 and 31st December 2019. For the CINAHL database, we limited the search to full-text research articles published from 1st January 2010 to 31st December 2019. For Science Direct, we defined the search to research articles published between 2010 and December 2019. Search targets the following keywords ("cellular phone" AND diet, "mobile phone" AND diet, smartphone AND diet, mHealth AND diet, iPads AND diet, "cellular phone" AND food, "mobile phone" AND food, smartphone AND food, mHealth AND food, iPads AND food, "cellular phone" AND nutrition, "mobile phone" AND nutrition, smartphone AND nutrition, mHealth AND nutrition, iPads AND nutrition ). The Boolean AND Figure 1. PRISMA flow chart of identification, screening, joins the two major parts. These yield 25950 results, reduced eligibility and inclusion of studies. to 13,897 after duplicates removal. They are screened based on titles, and we accessed a total of 775 articles for eligibility against our inclusion/exclusion criteria and study selection Evaluation Criteria and Data Extraction process. We scanned references of eligible studies to identify additional studies, but we have not included additional studies Reviewer GAT extracted all the selected studies’ key in this review. Finally, we included a total of 56 studies characteristics (study population, location, mobile app in this review process. Table 2 shows the search terms and details, and aim of the survey) shown in Table 3. Similarly, their corresponding search results. The PRISMA diagram in the Expert Group compromised of the authors of this Figure 1 shows the search flow and inclusion/exclusion of manuscript identified the attributes of each research question studies. shown in Table 4, 5, 6 and 7 that are mentioned in the existing literature of mHealth apps for data extraction to answer our Prepared using TRR.cls
4 Transportation Research Record XX(X) Table 2. Search Results (The Boolean AND joins the two major parts) Search Strings Search Results PUBMED CINAHL Science Direct "Cellular Phone" AND Diet 5 104 87 "Mobile Phone" AND Diet 51 622 1,057 "Mobile Telephone" AND Diet 2 106 79 Smartphone AND Diet 89 909 1,071 mHealth AND Diet 211 573 200 iPads AND Diet 6 287 263 "Cellular Phone" AND Food 1 180 400 "Mobile Phone" AND Food 30 1,018 4457 "Mobile Telephone" AND Food 1 132 314 Smartphone AND Food 64 1,396 3,418 mHealth AND Food 103 640 239 iPads AND Food 5 610 843 "Cellular Phone" AND Nutrition 2 136 106 "Mobile Phone" AND Nutrition 41 757 1,191 "Mobile Telephone" AND Nutrition 2 121 90 Smartphone AND Nutrition 90 1,030 1,139 mHealth AND Nutrition 183 663 171 iPads AND Nutrition 5 344 316 questions. Under the heading of general issues, we assessed Study Selection the difficulty in portion size estimation (79), demotivating The database search yielded 25,950 results. After removing information, dependence on expensive electrical devices such duplicates, we screened 13,897 based on titles. Out of that, as fit brands (50), the credibility of the database (78, 84) we excluded 13,122 studies, and 775 article texts were etc. Table S1 and S2 provide the extracted data of general assessed for eligibility by reviewers. Finally, we included issues in the supplementary material. We extracted the data 56 studies after excluding 719 text articles. The first study regarding stability of the application (82), usage reports included is from 2010. From 2013 to 2017, the publication (104), data confidentiality (85) and offline accessibility rate increased by 55.08%, with the highest number of studies (105) for back-end application issues. Table S3 provides published studies in 2017 (22.3%). We have the publication the extracted data in the supplementary material. For user year of all the included studies in our key characteristics interface requirements (28, 74–77), we extracted data of the table. attributes mentioned in Table 6. Table S4 and S5 in the In the subsequent sections, we have briefly described supplementary material provides the extracted data. Under mobile applications’ status as per our research questions. the heading of dietary components (78, 84), we extracted We have evaluated existing dietary applications by keeping the details of the attributes shown in Table 7. Table S6 in view critical features and general issues mentioned by and S7 in the supplementary material provides the extracted dietitians and users and supported by existing literature of data. We carried out the whole process by completing the mHealth apps. data extraction forms. Two researchers verified the data’s soundness and ensured data extracted from each study RQ1: What are the general problems that are justified the study’s aim. When the publications identified resolved by the freely available diet-related in the searches did not provide sufficient detail of mHealth apps, additional literature, websites, contacts with authors, or applications? application use itself was used to fill gaps. To determine the general issues found in existing applica- tions, we have categorized important parameters from differ- ent perspectives of users and dietitians as shown in Figure 2. Results For this purpose, we surveyed 35 freely available mHealth This section presents the results of the essential characteris- apps from included studies. tics of each selected study. It shows the results obtained from These general issues mentioned by dietitians and users in the extracted information to answer our research question. existing mHealth apps include credibility, (78) localization The brief detail of all the data extracted to answer our of database sources, (80), and difficulty in portion size research question is provided in the study’s supplementary estimation (79). Moreover, applications that require users material. to go through multiple steps for data entry (81) make the Prepared using TRR.cls
Ghalib Ahmed Tahir, Chu Kiong Loo, Nadine Kong and Foong Ming Moy 5 process cumbersome and negatively impact the whole user experience. (Figure 3) (A) shows the percentage breakdown of applications that looked to resolve these issues. Credibility of database sources is one of the major reasons for dietitians to not recommend apps to clients or patients due to concerns regarding their validity and questionable feedback in terms of accuracy. Nearly 34% of the applications (20, 21, 24, 29, 34, 36, 44, 48, 54, 55, 60, 62, 64) managed to resolve this issue by providing extensive details about the database, especially in terms of its sourcing. However, there are still some applications that offer little or no information about their database sources (17, 19, 27, 31, 33, 37, 45, 59, 68–70). Remaining applications (6, 15, 18, 22, 28, 32, 35, 41, 42, 65, 67) did not address the issue of credibility of database resources. On other hand only 10 out of 35 (6, 20, 21, 36, 41, 44, 45, 48, 55, 59) applications surveyed Figure 2. General issues faced by freely available diet-related had localized databases which are specific to certain region applications or culture. Subsequently, portion size estimation also plays a vital role when it comes to dietary applications. This component 59, 60, 62, 64, 65, 67–69). Frequent notifications often bother usually requires prior contextual knowledge to ensure better the user; especially when the app provides notifications after accuracy. Several apps deal with fixed food measurements nearly every step. Also, 40% of surveyed applications had no in terms of serving size, weight, or other simple household information available about notification or reminder settings, measurements. Generally, it is hard for most people to (6, 20, 24, 27, 31, 32, 35, 36, 42, 44, 48, 60, 68, 70) convert what they see on their plates to these measurements whereas remaining applications accommodated notification for entering into dietary apps. Moreover, when it comes system except “lose it” (18). Furthermore, all applications to Asian food, estimating portion size becomes even more except FoodWiz2 (62) and MyFitness Pal (Log2Lose) (67) challenging when multiple food items are mixed or placed do not require any sort of electrical device. Also, almost 37% on top of each other. Therefore, over or underestimating of applications (19) (21) (24) (27) (6) (31) (32) (44) (48) portion size is common for unskilled individuals, even failed to address the cumbersome process of entering meal more so in Asian families where each meal consists of information. In addition to this, a large number of existing multiple side dishes. These challenges make the estimation applications do not address the difficulty in portion size of portion size a complicated task for machine learning estimation. This is due to multidimensional challenges, from researchers, application developers, and dietitians. Many users’ perspective; as it is difficult to estimate the portion size existing applications (6, 20, 22, 27–29, 32, 34, 35, 48, 66–68) of food items used during preparation of food at different do not accommodate features which support estimation of restaurants. Moreover, lack of guidance/reference regarding portion size. Alternately, many existing applications (15, 17– the quantity, further complicates the procedure of estimating 20, 22, 27–29, 32–34, 36, 37, 41, 45, 54, 55, 59, 60, 62, 65– portion size. 70) require minimum steps for data recording ensuring a smooth user experience. RQ2: What are back-end application issues General issues faced by users involve frequent app crashes (82), cumbersome process of entering meal details, resolved by the freely available dietary demotivating information displays, dependence on expensive applications? electrical devices such as fit bands (50), frequent notifications The backend is an essential part of any mobile application, and difficulty in estimating portion size (80). As shown as it involves data storage, business logic, and security. in figure 3 (B), out of 35 surveyed applications, 6 Therefore, it plays the role of a server for mobile applications applications (41, 42, 44, 45, 69, 70) had no information about and stores information invisible to the end-users. Figure application’s stability in terms of app crashes. Both "happy" 4 presents backend application issues faced by end-users. (17) and "lose it" (18) experienced frequent app crashes. Backend issues generally involve no offline access to the key The remaining applications do not experience frequent app features (105), absence of usage reports (104), privacy or crashes. data confidentiality concerns (85), and frequent application Furthermore, large number of surveyed applications crashes (82). Therefore, the issues mentioned above should display motivating information and require less steps to be addressed by mobile applications to enhance the usability record user information (15, 17–22, 27–29, 32–37, 41, 54, 55, or user-friendliness of any diet-related application. Prepared using TRR.cls
6 Transportation Research Record XX(X) Figure 3. (A) and (B) shows percentage of applications that are resolving general problems faced by dietitians and users Reduced dependence on the Internet will not only improve Another important issue is lack of data confidentiality usability. It will also enhance the app’s responsiveness and privacy. Mobile applications, especially health-related and data processing, as the end-user is not restricted applications, should have concrete measures to ensure user’s from recording their data offline. Thus far, only 3 out of records’ confidentiality. Similarly, web services included in 35 surveyed applications allow offline accessibility (22) mobile applications should extract data without any leakages (41) (67), whereas other applications require an internet and minimal pilferage instances. Almost 22% of surveyed connection for data transmission to their respected servers. applications offer data privacy, while many studies do not Moreover, apps like Diet Cam rely on the client-server mention this problem. Moreover, some existing applications configuration for connectivity between mobile phones and such as Social POD (15), Happy (17), and ’mDPP’ (19) databases (21), which again requires a stable internet report user engagement or adherence to the app declined connection. over time. Figure 5 describes percentage of applications that managed to resolve aforementioned issues. RQ3: To what extent do the freely available dietary applications fulfill user interface requirements? Generally, user interface requirements encompass application Figure 4. Backend application issues design, user-friendliness, tutorial page, and two-user dash- board. As application design is one of the main requirements Similarly, technical bugs and frequent app crashes resulted (28, 76, 77), as it should be simple, have nice and appro- in unstable applications (17) (18)(19)(27)(6)(68)(54)(60). priate icons along with clear font size and color to improve Many of the applications surveyed suffer from technical usability. Out of 35 surveyed applications, 29 applications glitches and slow processing speed due to their dependence met the design criteria according to users’ requirements. on a reliable internet connection. Therefore, offline accessi- However, remaining applications (15, 19, 29, 35, 37, 48) did bility can help to address all these concerns. not endorse their design details. Prepared using TRR.cls
Ghalib Ahmed Tahir, Chu Kiong Loo, Nadine Kong and Foong Ming Moy 7 Figure 5. Shows percentage of applications that are resolving Backend application issues Another feature which enhances usability is the presence in punching in their updated information and updating their of appropriate tutorial pages (74, 75). For many diet- body weight regularly in the app, thereby keeping track of related applications, tutorial pages are preferred to show their progress. Similarly, alerts to consume meals at specified the metric measurements of serving size of food for a times, alerts for calories, drinking water, and doing exercises better understanding of the user. Unfortunately, most of the improve user engagement. 62% of total surveyed applications surveyed apps fail to provide this information, and only 20% (6, 16, 18–20, 23, 29, 32, 33, 35, 37, 41, 44, 45, 48, 69) were of surveyed applications (20, 22, 45, 55, 64, 67, 69) were able found out to provide such alerts to their users. Figure 6 shows to provide the details for a tutorial page. the percentage of existing applications that are implementing Two-user dashboards are another essential feature whereby these features. a simplified or easy-to-use version of the dashboard is available for patients or the general public. A more detailed RQ4: What are the dietary components and version is available for dietitians or researchers. Most critical features implemented by the freely surveyed applications do not incorporate this feature except for the Dietary Intake Assessment app (47). available dietary applications? Apart from these, user-friendliness has been qualified as Dietary component functionality mainly included evaluation the most important UI design component (77). Applications of diet quality, options to add supplements to the diet, are considered user-friendly when they have a complete history tracking, and storage of these records. Other essential data set, require fewer data entry steps, provide meaningful features include validity and comprehensiveness of database information, and have a user-friendly interface. Therefore, (78, 84), portion size estimation (79) and diet/nutrient bugs, glitches, and a cumbersome user interface of apps summary that provides information in terms of calories for can negatively influence the app’s usability. According to each meal as show in Figure 7. Figure 8 below illustrates the dietitians, almost 77% of the surveyed applications (6, 17, summary of the results gathered from surveyed applications. 18, 20–22, 24, 28, 32, 34–36, 41, 42, 44, 45, 48, 54, 55, To provide a good evaluation of diet quality, a dietary app 59, 60, 62, 64, 65, 67, 69, 70) were user friendly and had should display macronutrients’ balance and include reference higher rate of user engagement due to presence of simple values for interpretation. Based on this information, 34 out user interface and interactive design. According to the users, of 35 surveyed applications were able to assess diet quality 91% of the total applications surveyed were user-friendly. For properly. application design, most of the users prefer simple design and Moreover, options to track users’ weekly diet records easy-to-use apps. 70% of applications (17, 18, 20, 24, 27, 28, and their storage on websites for later use are considered 31, 32, 36, 41, 42, 45, 68) were attractive according to users’ important factors that can improve user experience. As per requirements. our survey total 26 applications(6, 15, 17, 18, 20–22, 28, 29, User interface requirements also involve information to 31, 32, 36, 44, 45, 48, 54, 55, 59, 60, 64, 65, 67–69) facilitate include in the user profile, and notification alerts to the user users by giving them access to their previous records. Items (76). Besides basic information, the application should allow included in existing applications should also be considered users to set goals in terms of desired body weight, and diet an essential feature, as some applications are particular about (76, 78). Only 53% of surveyed applications (15, 18, 23, 28, specific food items (like beverages). In contrast, other diet 30, 33, 35, 69) included goal-setting feature, while others did apps provide users with options to customize the food choices not even provide personalized profiles. accordingly. Finally, another important feature of diet-related applica- Another important feature which most of applications tions is notification alert to users (78). Reminders assist users (15, 17–19, 22, 24, 27, 28, 31, 34, 37, 41, 69) (21 out Prepared using TRR.cls
8 Transportation Research Record XX(X) Figure 6. Shows percentage of applications that are implementing key user interface features of 35 applications) failed to include, is the incorporation MyFitnessPal (22) display remaining calorie allowance to of a reliable and comprehensive food database. Apart from guide the user to achieve dietary goals. inclusion of database, validity of database also matters and unfortunately only 37% of the surveyed applications (6, 20, Furthermore, applications can make use of visual aids 21, 24, 29, 34, 36, 44, 48, 54, 55, 60, 62) possess validated by providing summaries of energy and nutrients intake in food database. the form of a diary, pie chart, table, and progress bar (77) Furthermore, applications should display the breakdown for better comprehension. Applications such as ENGAGED of nutrient components of the consumed food items. Data (34), provide goal thermometers to display user goals and should also include the proportion of calories of each meal the actual amount of calories and fat (in grams) consumed. (eg. Breakfast, lunch, dinner). However, some of the current Finally, we have ranked these applications based on the applications only include calories per meal, whereas other number of features they have implemented and mentioned in application like ‘Lose it’ (18) provides information about their study. Lark application, Food wiz2, Gocarb application balanced macronutrients. Another important functionality had a higher score in fulfilling the number of requirements brought forth by existing applications (6, 18, 21, 22, 24, from dietitians’ or researchers’ reference frames. On the other 24, 28, 29, 31, 36, 37, 42, 44, 45, 48, 70) is portion size hand, MyFitnessPal, Engaged, and MyFoodApp focuses estimation. Only few surveyed applications (21, 24, 24, 31, more on the general population’s requirements. Figure 10 and 36) use camera to estimate volume of a portion size, while Figure 9 show the applications score. most other surveyed apps rely on the standard household measurements (6, 18, 21, 22, 28, 29, 37, 42, 44? , 45). Water is considered an essential component of the human body, ensuring the proper functioning of multiple bodily functions. Therefore, it is equally as important to track users’ water intake. However, 14 out of 35 applications including Happy (17), Lose it (18), Metabolic Diet app (20), MyFitnessPal (22), and My Meal Mate (6) allows users to record total water intake. Whereas other diet-related applications tend to miss out on this important feature. As for the nutrient summary, most application databases include calorie information while other surveyed applications provide more specific nutrient information in databases. Apps like “Lose it” (18) provide information about three significant macronutrients like carbohydrates, proteins, and fats, which guide users to make better food choices. To enhance the user-friendliness of an application, tailored messages, feedback, and notifications according to user dietary intake (76) are essential factors to be considered. Figure 7. Important dietary component by users and dietitians Acting as guidelines for users, they improve the user- friendliness and user experience of an application. For instance, apps like My Meal Mate (6), ‘iDAT’ (32), and Prepared using TRR.cls
Ghalib Ahmed Tahir, Chu Kiong Loo, Nadine Kong and Foong Ming Moy 9 Figure 8. Shows percentage of applications that are implementing important dietary components Figure 9. Application score by keeping in view requirements from dietician perspective.Equal weight is given to each category and applications fulfilling more requirements have the highest score. The maximum total score is 4, and the max score of each category is one. Figure 10. Application score by keeping in view requirements from user perspective.Equal weight is given to each category and applications fulfilling more requirements have the highest score. The maximum total score is 4, and the max score of each category is one. Prepared using TRR.cls
10 Transportation Research Record XX(X) RQ5: What are the benefits and challenges 35 mHealth applications based on their usability, critical stemming from the included case studies? features and shows their strengths and weaknesses. The user-friendliness and high engagement are of considerable Current case studies have several benefits that can help users importance (77), especially since 68% of the existing monitor their daily diet and help them resolve diet-related mHealth apps incorporate this feature. We recommend that issues. Based on the shortlisted studies, we investigated user’s input in the development of mHealth interventions and freely available diet-related applications in terms of features, other considerations for end-users should be sought early on general problems, and usability challenges. Our findings aim in the process of app or digital health intervention design to to provide a broader view of current solutions to dietitians, ensure long and short term engagement (100) (101) (102) health experts, and researchers alike. Overall, the case studies (103). Similarly, the user notifications are equally important, equip both general users and health experts with information as it keeps them engaged and motivated (78). According to on the critical features that are not catered by most of the survey, we found that 62% of the apps provides timely the existing diet-related apps. Moreover, they will help alerts to the user. Likewise, goal setting also holds critical users quickly determine the viability of existing solutions significance, as it gives information about user’s personal to recommend further or use the solution that fulfills their preferences required for modification of their behaviour needs in the best possible way. Thus, these studies have paved accordingly (76, 78). Therefore, about 53% of the surveyed the way for the research community to introduce standard applications include the goal settings feature. guidelines for future diet-related apps according to criteria. As a result, the apps will be more substantial for patients, Likewise, our findings indicate that existing applications general users, and dietitians. are lagging in various aspects. Despite the importance of the credibility of database resources (78), only 30% Apart from this, the studies also highlight different of surveyed applications highlighted this issue. Besides challenges that can undermine current applications’ actual credibility, the comprehensive validation of the database with purpose. Major obstacles include integration and updating detailed information on macronutrients and micronutrients large food databases as food recipes, and their nutrient is also essential for clinical use. However, only 37% of content varies from region to region. In addition to this, the applications have included validated databases. Despite new food items are being introduced in the market every the rise of artificial intelligence, the methods for estimating day. Therefore, making the design and implementation portion size and logging food photos from the camera of such systems a difficult task. Similarly, a database’s have made significant advancements (38, 51), many apps comprehensiveness is one of the primary requisites of users to still depend on household measurements for portion size keep track of their micronutrients and macronutrients. Other estimation or manual entries of the food log. than this, the incorporation of user-friendliness along with notifications and personalized alerts are the challenges that Due to rising concerns of data security among users (85), the research community should consider. Also, data security diet-related apps must encrypt user data and use standardized is of vital importance (85) due to strict policies of regulating protocols to ensure data privacy and confidentiality. Yet, authorities and rising public concerns over sharing private the results indicate that only 22% of surveyed applications data. Therefore, diet-related apps should ensure data privacy have addressed data privacy issues. Similarly, there is a and confidentiality, which, unfortunately, many surveyed lack of economic data in existing studies to support using apps fail to address. mHealth apps for dietary assessment. Although the economic evaluation of mHealth apps is necessary to provide an evidence-based assessment of sustainability and benefits of Discussion investing in such technologies. (98) (99). We initiated this evaluation because of the rapid recent Despite considerations that existing diet-related apps emergence of freely available diet-related apps coupled with should address, all of the studies are valuable to broaden the increasing concerns over general issues, usability challenges research community’s knowledge. The identified applications and missingness of critical features. Following that, we in these works serve as a guide for users to choose between investigated the strength and weaknesses of freely available healthier alternatives and improve their dietary habits in the diet-related apps. The primary emphasis of previous reviews long term. by Rusin et al. (71), Kankanhalli et al. (72), and Prgomet Finally, we have made the following recommendations for et al. (73) was on functionalities and input methods or the research community based on our study. A localized the combined intervention of sleep and diet. Similarly, database is essential for nutritional assessment apps due Prgomet et al. (73) focused only on the inclusion of nutrition to variations in the food recipes and diet preferences information in the meal ordering system. among different cultures. Future diet-related apps should We focused on mHealth apps identified from existing also consider the technological advancement in artificial publications between 2010-2019. After carrying out the intelligence and explore the current methods of logging food literature search on three scientific databases, we evaluated and automated portion size estimation from food photos. Prepared using TRR.cls
Ghalib Ahmed Tahir, Chu Kiong Loo, Nadine Kong and Foong Ming Moy 11 It is noteworthy that several studies have implemented AI- Continual Lifelong Learning (IF0318M1006) from MESTECC, based strategies, but further investigation of these methods Malaysia and ONRG grant (Project No.: ONRG-NICOP- N62909- is required on a large scale. Furthermore, there is a dire 18-1-2086) / IF017-2018 from Office of Naval and Research Global, need to develop standard guidelines for the development UK. of diet-related apps, as standardized solutions will be more reliable in the future for patients, general users, and dietitians. References Finally, when designing modern diet-related applications, the research community should consider our findings to enhance 1. “Obesity and Overweight.” World Health Organization the usability and completeness of the solution. (WHO) 2. P. Farago, "iOS and Android Adoption Explodes Internation- Limitations of the data gathered for this study ally," The Flurry Blog, 2013. 3. D. Amitava and P. Abhinay and R. Rahul and S. Priya, This review has limitations that require further investigations. "Technology diffusion: Shift happens — The case of iOS Firstly, the analysis was limited to studies published in the and Android handsets," Technological Forecasting and Social searched databases and only written in the English language. Change, vol. 118, p. 28–43, 2017. Related articles in other languages were not included. 4. Y. Hairong and H. Hongwei and X. Youzhi and G. Secondly, this research does not consider demographic Mikael, "Wireless sensor network based E-health sys- information about a particular race or culture while designing tem—implementation and experimental results," IEEE Trans- the research questions. actions on Consumer Electronics , vol. 56, p. 2288–2295, 2010. Conclusion 5. H. Alfredo and M. Fernando and V. Guillermo and P. Gianfranco and C. Guy, "Real-time ECG transmission via Dietary apps for nutritional assessment are developed to internet for nonclinical applications.," IEEE Transactions on assist users with their diet-related issues or keep track of Information Technology in Biomedicine , vol. 5, p. 253–257, their dietary intake. Such apps tend to act as guides and 2001. enable users to choose healthier alternatives to improve 6. C. Michelle and B. Victoria and C. Janet., "Development their nutritional habits in the long term. Therefore, due of ’My Meal Mate’ - A smartphone intervention for weight to the vital importance of diet-related apps, this SLR loss.," Nutrition Bulletin, vol. 38, pp. 80-84, 2013. analyzed a wide range of existing literature on mHealth 7. D. BL and T.Schap and R. Ettienne-Gittens and F. Zhu et al. apps from scientific databases of CINAHL, Science Direct, , "Novel technologies for assessing dietary intake: evaluating and PUBMED and shortlisted almost 56 studies. We have the usability of a mobile telephone food record among adults investigated the apps’ comprehensiveness in terms of critical and adolescents," J Med Internet Res, vol. 14, p. 58, 2012. features, general issues, and usability challenges from general 8. S. Lim and P.Bentley and N.Kanakam and F.Ishikawa and users’ reference frames. We have further examined the S.Honiden., "Investigating country differences in mobile app strength and weaknesses of the existing freely available diet- user behavior and challenges for software engineering," IEEE related apps and summarized concerns and gaps for future Transactions on Software Engineering, vol. 41, p. 40–64, work. Our findings show that the credibility of database 2015. resources, comprehensive information about macronutrients 9. K.Hsu,., "Efficiently and Effectively Mining Time- and micronutrients, validation of database, data privacy, use Constrained Sequential Patterns of Smartphone Application of AI for food logs, and automated portion size estimation Usage," Mobile Information Systems, 2017. from the pictures are foremost challenges. Addressing the 10. E. Din, "Ergonomic requirements for office work with visual challenges mentioned above will improve the usability and display terminals (VDTs)–Part 11: Guidance on usability, comprehensiveness of diet-related apps. Therefore, making International Organization for Standardization," 1998. them more substantial for patients, general users, and 11. J. Nielsen, "Usability 101:Introduction to usability," 2003. dieticians. Moreover, implementing blockchain technology 12. A.Luca and S.Das and M.Ortlieb and I.Ion and B.Laurie, and health standards for data security, exploring recent trends "Expert and Non-Expert Attitudes towards (Secure) Instant in continual learning for food recognition, and outlining Messaging,” Symposium on Usable Privacy and Security standard guidelines for regulating apps are essential future (SOUPS)," 2016. topics that can be explored. 13. L. Carvajal and A. Moreno and M.Sanchez-Segura and A. Seffah., "Usability through software design," IEEE Acknowledgements Transactions on Software Engineering, vol. 39, no. 11, p. This research was supported by the UM Partnership Grant: 1582–1596, 2013. Project No: RK012-2019 from University of Malaya, IIRG Grant 14. D.Moher and A. Liberati and J. Tetzlaff and D.G. Altman, , (IIRG002C-19HWB) from University of Malaya, International "Preferred Reporting Items for Systematic Reviews and Meta- Collaboration Fund for project Developmental Cognitive Robot with Analyses: The PRISMA Statement," plos one medicine, 2009. Prepared using TRR.cls
12 Transportation Research Record XX(X) 15. S. Hales and GM. Turner-McGrievy and S. Wilcox , et al., in Australian truck drivers," Journal of Science and Medicine "Social networks for improving healthy weight loss behaviors in Sport, vol. 18, p. 124, 2014. for overweight and obese adults: A randomized clinical trial 27. H. Holmen and A.Torbjørnsen and A. Wahl and A. Jenum and of the social pounds off digitally (Social POD) mobile app.," M. Cvancarova and E. Arsand and L. Ribu., "A Mobile Health Int J Med Inform., vol. 94, pp. 81-90, 2016. Intervention for Self-Management and Lifestyle Change for 16. M. Mccarroll and S. Armbruster and R. Pohle-Krauza and Persons With Type 2 Diabetes, Part 2: One-Year Results From A. Lyzen and S. Min and D. Nash and G. Roulette and the Norwegian Randomized Controlled Trial RENEWING S. Andrews and V. Gruenigen, " Feasibility of a lifestyle HEALTH.," JMIR Mhealth Uhealth, vol. 2, p. 57, 2014. intervention for overweight/obese endometrial and breast 28. S. Gabrielli and M. Dianti and R. Maimone and M. Betta and cancer survivors using an interactive mobile application.," L. Filippi and M. Ghezzi and S. Forti., "Design of a Mobile Gynecol Oncol. , vol. 137, p. 508–515, 2015. App for Nutrition Education (TreC-LifeStyle) and Formative 17. N. Ribeiro and L. Moreira and AM. Almeida and F. Santos- Evaluation With Families of Overweight Children," JMIR Silva, "Happy: Cancer Prevention Using Smartphones," Cancer, vol. 5, p. 48, 2017. Procedia Computer Science, vol. 100, p. 466 – 473 , 2016. 29. W. Lippevelde and J. Vangeel and N. De Cock and C. Lachat 18. L. Burke and Y. Zheng and Q. Ma and J. Mancino and I. Loar and L. Goossens and K. Beullens and L. Vervoort and C. and E. Music and M. Styn and L. Ewing and B. French and D. Braet and L. Maes and S. Eggermont and B. Deforche and Sieworek and A. Smailagic and S. Sereika., "The SMARTER J. Camp (2016). "Using a gamified monitoring app to change pilot study: Testing feasibility of real-time feedback fordietary adolescents’ snack intake: the development of the REWARD self-monitoring," Preventive Medicine Reports, vol. 6, pp. app and evaluation design.," BMC Public Health, 2016. 278-285, 2017. 30. M. Carter and V. Burley and J. Cade, "Weight Loss Associated 19. Y. Fukuoka and C. Gay and K.Joiner and E. Vittinghoff, "A With Different Patterns of Self-Monitoring Using the Mobile Novel Diabetes Prevention Intervention Using a Mobile App," Phone App My Meal Mate," JMIR mHealth and uHealth, American journal of preventive medicine, vol. 49, pp. 223- 2017. 237, 2015. 31. SL. Casperson and J. Sieling and J. Moon and L. Johnson and 20. G. Ho and Ueda, Keiko and Houben, Roderick and Joa, Jeff JN. Roemmich and L. Whigham, "A mobile phone food record and Giezen, Alette and Cheng, Barbara and Karnebeek, Clara., app to digitally capture dietary intake for adolescents in a free- "Metabolic Diet App Suite for inborn errors of amino acid living environment: usability study," JMIR Mhealth Uhealth, metabolism," Molecular Genetics and Metabolism, vol. 117, vol. 3, p. 30, 2015. no. 3, pp. 322-327, 2016. 32. G. Goh and NC. Tan and R. Malhotra et al. "Short-term 21. Fanyu Kong, Jindong Tan, DietCam: Automatic dietary trajectories of use of a caloric-monitoring mobile phone app assessment with mobile camera phones, Pervasive and Mobile among patients with type 2 diabetes mellitus in a primary care Computing,Volume 8, Issue 1, 2012,Pages 147-163,SSN setting," J Med Internet Res, pp. 33-33, 2015. 1574-1192, https://doi.org/10.1016/j.pmcj.2011.07.003. 33. DA. Kerr and CM. Pollard and P. Howat, et al. "Connecting 22. C. Levinson and L. Fewell and L. Brosof., "My Fitness Pal Health and Technology (CHAT):protocol of a randomized calorie tracker usage in the eating disorders," Eat Behav., controlled trial to improve nutrition behaviours using mobile 2017. devices and tailored text messaging in young adults," BMC 23. C. Wharton and C. Johnston and B. Cunningham and D. Public Health, 2012. Sterner, "Dietary self-monitoring, but not dietary quality, 34. C. Pellegrini and J. Duncan and A. Moller and J. Buscemi and improves with use of smartphone app technology in an 8-week A. Sularz and A. DeMott and A. Pictor and S. Pagoto and J. weight loss trial," J Nutr Educ Behav, vol. 46, 2014. Siddique and B. Spring, "A smartphone-supported weight loss 24. BL. Six and TE. Schap and FM. Zhu and A. Mariappan program: design of the ENGAGED randomized controlled and M. Bosch and EJ. Delp and DS. Ebert and DA. Kerr trial," BMC Public Health, vol. 12, 2012. and CJ. Boushey, " Evidence-based development of a mobile 35. J. Recio-Rodríguez et al.,"Short-Term Effectiveness of a telephone food record," Journal of the American Dietetic Mobile Phone App for Increasing Physical Activity and Association, vol. 110, p. 74–79, 2010. Adherence to the Mediterranean Diet in Primary Care: A 25. Six BL, Schap TE, Kerr DA, Boushey CJ. Evaluation of the Randomized Controlled Trial (EVIDENT II Study).," J Med Food and Nutrient Database for Dietary Studies for use with Internet Res, vol. 12, pp. 1-1, 2016. a mobile telephone food record. J Food Compost Anal. 2011 36. D. Rhyner and H. Loher and J. Dehais and M. Anthimopoulos Dec 1;24(8):1160-1167. doi: 10.1016/j.jfca.2011.06.006. and S. Shevchik and R. Botwey and D. Duke and C. Stettler PMID: 22389554; PMCID: PMC3289151. and P. Diem and S. Mougiakakou., "Carbohydrate Estimation 26. N. Gilson and T. Pavey and S. Gomersall and C. Vandelanotte by a Mobile Phone-Based System Versus Self-Estimations of and M. Duncan and O. Wright and S.Trost and W. Brown. , Individuals With Type 1 Diabetes Mellitus: A Comparative "Shifting gears: Process evaluation of an activity tracker and Study.," JMIR Mhealth Uhealth, vol. 18, p. 101, 2016. smart phone application to promote healthy lifestyle choices Prepared using TRR.cls
Ghalib Ahmed Tahir, Chu Kiong Loo, Nadine Kong and Foong Ming Moy 13 37. E. Seto and J. Hua and L.Wu and V.Shia and S. Eom et al., Mobile Phone Dietary Assessment App," JMIR Mhealth "Models of Individual Dietary Behavior Based on Smartphone Uhealth, vol. 4, p. 92, 2016. Data: The Influence of Routine, Physical Activity, Emotion, 49. C. Martin and A. Gilmore and J. Apolzan and C. Myers and and Food Environment," PLOS One, 2016. D. Thomas and L. Redman. , "Smartloss: A Personalized 38. F. Zhu and M. Bosch and C. Boushey and E. Delp, "AN Mobile Health Intervention for Weight Management and IMAGE ANALYSIS SYSTEM FOR DIETARY ASSESS- Health Promotion.," JMIR Mhealth Uhealth, vol. 4, p. 18, MENT AND EVALUATION.," Proc Int Conf Image Proc:, pp. 2016. 1853-1856, 2010. 50. E. Arsand and M. Muzny and M. Bradway and J. Muzik 39. Y. Probst and DT. Nguyen and M. Tran and W. Li., "Dietary and G. Hartvigsen., "Performance of the first combined Assessment on a Mobile Phone Using Image Processing smartwatch and smartphone diabetes diary application study," and Pattern Recognition Techniques: Algorithm Design and JMIR Mhealth Uhealth, vol. 9, pp. 556-563, 2015. System Prototyping," Nutrients, vol. 7, no. 8, pp. 6128-6138, 51. F. Zhu and M. Bosch and I. Woo and S. Kim and C. Boushey 2015. and D. Ebert and E. Delp. , "The Use of Mobile Devices in 40. K. Serrano and M. Yu and K. Coa and L. Collins and A. Aiding Dietary Assessment and Evaluation," IEEE J Sel Top Atienza, . "Mining Health App Data to Find More and Less Signal Process, vol. 4, pp. 756-766, 2010. Successful Weight Loss Subgroups," J Med Internet Res, vol. 52. Ali ZC, Silvioli R, Rajai A, Aslam TM. Feasibility of Use 18, p. 60, 2016. of a Mobile Application for Nutrition Assessment Pertinent 41. C. Farsjø and A. Moen, "New app can give nutritional support to Age-Related Macular Degeneration (MANAGER2). Transl to home-dwelling elderly," Norwegian Journal of Clinical Vis Sci Technol. 2017;6(1):4. Published 2017 Jan 20. Nursing / Sykepleien Forskning, pp. 1-10, 2017. doi:10.1167/tvst.6.1.4 42. I.De La Torre Díez and B. Zapirain and M. López-Coronado 53. Ahmed M, Mandic I, Lou W, Goodman L, Jacobs I, L’Abbé and J. Rodrigues and C. Vegas., "A New mHealth App for MR. Validation of a Tablet Application for Assessing Dietary Monitoring and Awareness of Healthy Eating: Development Intakes Compared with the Measured Food Intake/Food Waste and User Evaluation by Spanish Users.," J Med Syst, vol. 41, Method in Military Personnel Consuming Field Rations. pp. 1-7, 2017. Nutrients. 2017 Feb 27;9(3):200. doi: 10.3390/nu9030200. 43. E. Delisle Nyström Forsum and H, Henriksson and Y. PMID: 28264428; PMCID: PMC5372863. Lagerros and C. Larsson and R. Maddison and T. Timpka and 54. G. Ambrosini and M. Hurworth and R. Giglia and G. Trapp M. Löf, "A Mobile Phone Based Method to Assess Energy and and P. Strauss., Feasibility of a commercial smartphone Food Intake in Young Children: A Validation Study against the application for dietary assessment in epidemiological research Doubly Labelled Water Method and 24 h Dietary Recalls.," and comparison with 24-h dietary recalls," Nutrition Journal, Nutrients, vol. 8, pp. 1-11, 2016. vol. 17, 2017. 44. A. Rangan and L. Tieleman and JCY. Louie and LM. Tang 55. M. Bardus and G.Hamadeh and B. Hayek and R. Al Kherfan,. and L.Kairey and R. Roy and J. Kay and M. Allman- , "A Self-Directed Mobile Intervention (WaznApp) to Promote Farinelli., "Electronic Dietary Intake Assessment (e-DIA): Weight Control Among Employees at a Lebanese University: relative validity of a mobile phone application to measure Protocol for a Feasibility Pilot Randomized Controlled Trial," intake of food groups," British Journal of Nutrition, vol. 1, pp. JMIR Research Protocols, vol. 7, 2018. 2219-2226, 2016. 56. G. Bennett, "Effectiveness of an App and Provider Counseling 45. S. Mummah and T. Robinson and M. Mathur and S. for Obesity Treatment in Primary Care," American Journal of Farzinkhou and S. Sutton and C. Gardner., "Effect of a mobile Preventive Medicine, vol. 55, 2018. app intervention on vegetable consumption in overweight 57. J. Chen and W. Berkman and M. Bardouh and C.Ng and adults: a randomized controlled trial.," Int J Behav Nutr Phys M. Allman-Farinelli, "The use of a food logging app in the Act, vol. 14, p. 125, 2017. naturalistic setting fails to provide accurate measurements of 46. S. Mummah and A. King and C. Gardner and S. Sutton., nutrients and poses usability challenges," pp. 208-216, 2019. "Iterative development of Vegethon: a theory-based mobile 58. N. De Cock and J. Vangeel and C. Lachat and K. Beullens and app intervention to increase vegetable consumption.," Int J L. Vervoort and L. Goossens and L. Maes and B. Deforche and Behav Nutr Phys Act, vol. 13, p. 90, 2016. S. Henauw and C. Braet and S. Eggermont and P. Kolsteren 47. P. Hull and J.Emerson and M. Quirk and J. Canedo and J. and J. Camp and W. Lippevelde, "Use of Fitness and Nutrition Jones and V. Vylegzhanina and D. Schmidt and S. Mulvaney Apps: Associations With Body Mass Index, Snacking, and and B. Beech and C. Briley and C. Harris and B. Husaini., "A Drinking Habits in Adolescents," JMIR mHealth and uHealth, Smartphone App for Families With Preschool-Aged Children vol. 5, p. e58, 2017. in a Public Nutrition Program: Prototype Development and 59. E. Everett and B. Kane and A. Yoo and A. Dobs and N. Beta-Testing," JMIR Mhealth Uhealth, vol. 5, p. 102, 2017. Mathioudakis. "A Novel Approach for Fully Automated, 48. A. Svensson and M. Magnusson and C. Larsson and, Personalized Health Coaching for Adults with Prediabetes: "Overcoming Barriers: Adolescents’ Experiences Using a Pilot Clinical Trial," vol. 20, no. 2, 2018. Prepared using TRR.cls
You can also read